Central Bank Independence (CBI) and Exchange Rate: Does the CBI Shock Matter? - IEA ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Central Bank Independence (CBI) and Exchange Rate:
Does the CBI Shock Matter?
Cep Jandi Anwar
Department of Economics, Faculty of Economics and Business,
University of Sultan Ageng Tirtayasa, Indonesia.
E-mail: cepjandianwar@untirta.ac.id
Abstract
In this study, we estimate the responses of exchange rate and economic activ-
ity in relation to CBI shocks in developing countries. The analysis is based on
a panel Vector Autoregressive (Panel VAR) estimation. By applying poola-
bility tests, we identify heterogeneity across the countries in our sample. One
possible solution to this problem is to apply a mean-group estimation to the
panel VAR. Additionally, we divide our sample countries to make the sub-
group poolable. The results reveal that CBI requires up to a year to cause
an appreciation in the exchange rate. We also conclude that exchange rate
has an essential role to play in monetary policy transmission, to the extent
that a change in CBI affects the exchange rate, thereby influencing private
consumption and investment.
Keywords:
CBI, Exchange Rate, Economic Activity, Panel VAR, Poolability Test,
Mean-Group Estimation.
JEL Classifications: E20, E58, F31
Preprint submitted to The 19th IEAWC January 7, 20201. Introduction
A major research has succeeded to reveal that central bank independence
(hereafter, CBI) is a main factor on price stability (Cukierman et al., 1992;
Acemoglu et al., 2008; Agoba et al., 2017). However, according to Bearce
(2008), assign a price stability as a main objective of CBI may have a nega-
tive consequence to exchange rate stability. The main reason of this negative
effect is because under international capital mobility, policymaker encoun-
ters a trade off between domestic monetary independence and exchange rate
stability. In short, policymakers have only one monetary policy, which can
either be directed at nominal external stability or towards a domestic policy
goal, including growth and inflation. Moreover, Government in developing
countries has ability to manipulate exchange rate depreciation to win the
election. However, if the central bank is fully credible and free from govern-
ment interference, this may limit the government on such manipulation.
Developing countries should pay more attention regarding exchange rate
depreciation because it may have impact on macroeconomic stability. Some
countries in Asia, such as Indonesia, Thailand and Korean experienced fi-
nancial crisis in 1997 due to depreciation of the exchange rate. Thus, those
countries increased their credibility and autonomy their central banks to cre-
ate stable domestic currencies.
Furthermore, exchange rate play an important role in macroeconomic
stability because the exchange rate is an essential factor affecting exports
and imports price. Thus, exchange rate influence imported goods prices
in domestic market and domestic goods in foreign market. Consequently,
changes in exchange rate may lead to significant effect on domestic inflation,
particularly in small open countries, and finally affecting public consumption
and investment.
This paper seeks to contribute to the empirical research by linking the
relationship between CBI, exchange rate and economic activity in developing
countries. In this paper, we examine the interaction between CBI, exchange
rate and economic activity by fitting a panel VAR estimation on quarterly
data spanning the periods 1991Q1 and 2016Q4. Subsequently, we test the
pooling assumption of the model by applying the Chow and Roy-Zellner tests.
We establish that the models contain heterogeneity among the samples; thus,
we apply a mean-group estimation for the panel VAR by averaging all of the
individual VAR coefficients. We also divide our samples into three subgroups
to make a poolable group. Subsequently, we can compare the subsamples and
2the full sample in regard to the link between CBI, exchange rate, consumption
and investment.
The rest of this study is organised as follow. Section 2 provides a related
review. Section 3 explores the data set, construction methodology used and
models. Section 4 discusses the empirical results. Section 5 is the concluding
session.
2. Related Literature
2.1. Theoretical Review on CBI and Exchange Rate
We develop a model similar to Sanchez (2008) in order to investigate the
relationship between CBI and exchange rate. In this context, the behaviour
of the private sector can be described as follows:
πt = Et−1 πt + α(yt − εst ) − γ(et − Et−1 et ) (1)
yt = −βrt − δet + εdt (2)
rt = −Et et+1 + et + εft (3)
rt = Rt + Et πt+1 (4)
where all variables (except the interest rates) are in term of logarithms
and constants have been normalised to zero. All parameters are assumed
to be positive, except δ which can apply for any real value. The negative δ
denotes a contractionary depreciation, while the positive means an expan-
sionary depreciation. Equation 1 is augmented Phillips curve, where inflation
rate (π) is determined by expectation of inflation (Et−1 π), output gap (y)
and exchange rate (e) pass through term. A higher e denotes an apprecia-
tion. Expression 2 denotes that output depends on the real interest rate (r),
real exchange rate and demand shock (εdt ). Equation 3 states an uncovered
interest parity condition. Equation 4 is the Fisher equation. εst and εdt are
i.i.d. white noise shock with mean zero, whilst εft denotes an autocorrelated
disturbance.
According to Rogoff (1985), the central bank minimises the following loss
3function:
1
L = E[χ(πt − π̃t )2 + α(yt − εst )2 ] (5)
2
It is also assumed that the public knows α, β, δ and the distribution of
disturbances (εst , εdt and εft ) and that it observes the nominal interest rate
and nominal exchange rate. The central bank concerns about the deviation
of inflation from its target, πt − π̃t , and the output gap, yt − εst .
From Equation 1 we have:
α(yt − εst ) = (πt − Et−1 πt ) + γ(et − Et−1 et ) (6)
Substituting Equation 6 to 5:
1
L = E[χ(πt − π̃t )2 + (πt − Et−1 πt ) + γ(et − Et−1 et )2 ] (7)
2
Differentiating Equation 7, with respect to π, we obtain:
χπt − χπ̃t + πt − Et−1 πt + γ(et − Et−1 et ) = 0 (8)
(1 + χ)πt = Et−1 πt − γ(et − Et−1 et ) + χπ̃t
1 χ
πt = [Et−1 πt − γ(et − Et−1 et )] + π̃t
(1 + χ) (1 + χ)
Thus, we have an optimal policy of inflation under discretion:
χ χ
πt = 1 − [Et−1 πt − γ(et − Et−1 et )] + π̃t (9)
(1 + χ) (1 + χ)
Assuming rational expectation which expected inflation equals expected
4targeted inflation, we have
Et−1 πt = Et−1 π̃t (10)
Substituting Equation 10 to Equation 9, we get the optimal inflation rate,
πtopt :
χ χ χ
πtopt =− 1− γ(et − Et−1 et ) + π̃t + 1 − Et−1 πt
(1 + χ) (1 + χ) (1 + χ)
(11)
The optimal inflation rate captures the effect of unexpected exchange rate
fluctuation on prices and the weighted expectation of the inflation target and
the actual inflation target.
After getting an optimal inflation rate, we then find the optimal interest
rate. First, combining Equations 1 and 2, then we have:
1 δ 1 γ 1
rt = Et−1 πt − et + εxd
t − (et − Et−1 et ) − πt (12)
αβ β β αβ αβ
Second, substituting Equation 12 to Equation 11, we get the optimal
interest rate, rtopt :
1 xd χ χγ δ
rtopt = εt − (π̃t − Et−1 π̃t ) − (et − Et−1 et ) − et
β αβ(1 + χ) αβ(1 + χ) β
(13)
where εxdt ≡ εdt − εst . The optimal central bank reaction function in
Equation 10 implies that the interest rate should be increased to balance
positive unexpected excess demand pressure, while that it should be reduced
if the inflation target is relaxed or the exchange rate appreciated.
By assuming that the inflation target applies a fixed and credible value
5of π̃t ). From Equation 3, we have:
et = rt + Et et+1 et − εft (14)
Combining Equations 13 and 14, we have:
1 xd χ χγ δ
et = εt − (π̃t − Et−1 π̃t ) − (et − Et−1 et ) − et
β αβ(1 + χ) αβ(1 + χ) β
+ Et et+1 et − εft (15)
δ 1 χ χγ
et (1 + ) = εxd
t − (π̃t − Et−1 π̃t ) − (et − Et−1 et )
β β αβ(1 + χ) αβ(1 + χ)
+ Et et+1 et − εft (16)
β+δ 1 χ χγ
et ( ) = εxd
t − (π̃t − Et−1 π̃t ) − (et − Et−1 et )
β β αβ(1 + χ) αβ(1 + χ)
+ Et et+1 − εft (17)
1 xd χγ β β f
et = εt − (et − Et−1 et ) + Et et+1 − ε
β+δ (β + δ)(1 + χ)α β+δ β+δ t
(18)
6Taking the first derivative of et with respect to χ, we can find:
∂et γ 1
=− (et − Et−1 et ) < 0 (19)
∂χ α(β + δ) (1 + χ)2
The higher CBI, χ, generates a higher exchange rate. In other words,
more independent central banks are more likely to appreciate the exchange
rate.
2.2. Empirical Literature
The effect of monetary policy on exchange rate has become one of the
most interesting research in macroeconomic policy since the last two decades.
Moser and Dreher (2010) examined the effect of changing the governor of the
central bank on the foreign exchange market, domestic stock market and
sovereign bond spreads based on a data set for 20 emerging countries over
the period 1992 to 2006. Consequently, asset price should change to the
extent of their sensitivity to inflation. Their results show that changing the
governor of the central bank has a negative effect on the financial market.
The reasons why investors respond negatively is because the new governor
of the central bank apparently suffers from a systemic credibility problem.
Eichler and Littke (2018) investigated the effect of CBI on exchange rate
volatility using panel data for 62 economies from 1998 to 2010. They reported
that a conservative and independent central bank will reduce the public’s un-
certainty about the central bank’s policy objective, thus reduce the volatility
of inflation expectation, and finally creating a pronounced stabilising effect
on exchange rate volatility. They also revealed that exchange rate volatility
depends on price flexibility in the goods market, central bank preferences
for price stability and the interest rate sensitivity related to money demand.
They established strong empirical evidence that an increase in independent
central banks decreases exchange rate volatility.
The relationship between asset price and economic activity is vary de-
pending on financial structure. Gerlach and Assenmacher-Wesche (2008)
studied the response of asset prices (property price and equity price), infla-
tion and real activity to interest rate policy shocks for a panel of 17 OECD
countries using quarterly data for the period 1986 to 2006. By way of per-
formed individual VAR, they discovered that inflation and output growth
respond better to interest rate shock than property and equity shock. Fur-
7thermore, their study applies a mean group estimator for panel VAR, which
split the sample into two groups based on financial structure. They con-
cluded that the interest rate shock has a positive effect on housing price and
increase real economic activity. However, for equity price, monetary policy
shock reduces the equity price. Panel VAR analysis of different subgroups
of countries reveals that the effect of interest rate on housing price seems
influenced by financial structure.
3. Econometric Methodology
The primary goal of this study is to observe the impact of CBI shock
on asset prices in developing countries. To solve this problem, we apply a
panel VAR proposed by Canova and Ciccarelli (2013). In the panel VAR
model, all variables are considered as endogenous and interdependent but a
cross-sectional dimension is included in the representation. Let, Yt as the
stacked model of yit , the vector of G variables for each unit i=1,...,N , i.e.,
Yt = (y01t , y02t , ..., y0N t )0 where i is generic and indicate countries. Then, a panel
VAR is:
yit = A0i (t) + Ai (`)Yt−1 + uit i = 1, ..., N t = 1, ..., T (20)
where A(`) is a polynomial in the lag operator and iid, A0 t is the deterministic
components, uit is a G × 1 vector of random disturbances. Equation (20)
may include constants, seasonal dummies and deterministic polynomial in
time.
A typical variation of (20) allows the G variables in Yt to be linear function
of a set of predetermined or exogenous variables, Wt . Then Equation (20)
can be written as:
yit = A0i (t) + Ai (`)Y1t−1 + Fi (`)Wt + uit (21)
where ut = [u1t , u2t , ..., uN t ]0 ∼ iid(0, ), Fi,j are G × M matrices for each
P
lag j = 1, ..., q, and Wt is a M × 1 vector of predetermined or exogenous
variables, common to all unit i.
A panel VAR has three characteristic features. First, lags of all endoge-
nous variables of all units enter the model for unit i, it is known by dynamic
8interdependencies. Second, uit is generally correlated across i, it is called
static interdependencies. The third characteristic is cross-section hetero-
geneity where the intercept, the slope, and the variance of the shocks u1it
may be unit specific.
Our Panel VAR model to analyse the interaction among CBI, exchange
rate and economic activity includes four endogenous variables: exchange rate
(ER), CBI, household’s consumption (Cons), and investment (Inv). Turning
now to the full panel VAR case, in our model G=4, thus our Panel VAR
models are:
k
X k
X k
X k
X
ERit = α1,i + a1,j ERi,t−j + b1,j CBIi,t−j + c1,j Consi,t−j + d1,j Invi,t−j + u1,it
j=1 j=1 j=1 j=1
(22a)
k
X k
X k
X k
X
CBIit = α2,i + a2,j ERi,t−j + b2,j CBIi,t−j + c2,j Consi,t−j + d2,j Invi,t−j + u2,it
j=1 j=1 j=1 j=1
(22b)
k
X k
X k
X k
X
Consit = α3,i + a3,j ERi,t−j + b3,j CBIi,t−j + c3,j Consi,t−j + d3,j Invi,t−j + u3,it
j=1 j=1 j=1 j=1
(22c)
k
X k
X k
X k
X
Invit = α4,i + a4,j ERi,t−j + b4,j CBIi,t−j + c4,j Consi,t−j + d4,j Invi,t−j + u4,it
j=1 j=1 j=1 j=1
(22d)
4. Data
The panel data used in this model covers 26 developing countries1 . Our
dataset consists of 4 variables: CBI, exchange rate, consumption and invest-
ment. Quarterly data from the years 1991 quarter 1 to 2016 quarter 4 are
used.
For the measure of CBI, we follow the CBI index constructed by Cukier-
man et al. (1992). This index is based on the legal aspect of independence.
The index is between 0 and 1, with higher values denoting greater CBI for
1
Argentina, Bolivia, Egypt, Ethiopia, Ghana, Guatemala, Honduras, Indonesia, Kenya,
Malaysia, Mauritania, Mexico, Morocco, Nicaragua, Pakistan, Paraguay, Philippines,
South Africa, Suriname, Thailand, Trinidad and Tobago, Tunisia, Turkey, Uruguay,
Venezuela and Zambia
9the legal index. The data relating to the CBI index is legal variable aggregate
weighted obtained from Garriga’s (2016) data set.
The role of asset prices is represented by the exchange rate. We use the
exchange rate in terms of logarithm natural. Exchange rate is the bilateral
currency of each country’s sample against the U.S. dollar (USD). The data are
retrieved from the International Financial Statistics (IFS) of the International
Monetary Fund (IMF). We use household consumption and investment as a
real economic activity following Claessens and Kose (2017). In terms of
logarithm natural, the data are retrieved from the IFS of the IMF.
5. Empirical Findings
5.1. Full Sample Countries Panel VAR
First, we examine the interrelationship between CBI, exchange rate and
economic activity by applying a panel VAR. Lag 2 is selected as the optimal
lag based on the Akaike information criterion. The panel VAR models in
Equations (22a) - (22d) are estimated in pooled least squared (POLS). The
POLS estimator is known to be potentially biased in a dynamic panel setting
if the coefficients on the endogenous variables differ across-countries. We run
the Chow and Roy-Zellner tests proposed by Baltagi (2008) to investigate the
heterogeneity coefficients in the model. The null hypothesis confirms that
the coefficients are the same for all country sample. While the alternative
hypothesis is the coefficients differ all cross-countries. The Chow and Roy-
Zellner tests show that we reject the null hypothesis. This result implies that
coefficients in the panel VAR model contain cross country heterogeneity. One
common way to solve the heterogeneity problem in the model is to perform
the mean-group estimation proposed by Pesaran and Smith (1995), which has
also been used in previous studies, such as those of Gerlach and Assenmacher-
Wesche (2008) and Sa et al. (2011).
It is difficult to interpret the coefficient of VAR; hence, we focus the
analysis on Impulse Response Function (IRF) and variance decomposition.
In Figure 1, we display the impulse responses functions over 20 quarters for
a one standard deviation shock implied by the panel VAR regression using
the mean-group estimator. We will focus on the exchange rate response
to CBI shock, consumption and investment responses to the exchange rate,
consumption and investment responses to CBI shocks. The response of the
exchange rate to one standard deviation shock of CBI is positive in the
beginning, after the shock. This finding indicates that in the short-run, the
10Figure 1: Impulse Response Function Mean Group Estimation Model 1
CBI shock depreciates the exchange rate. After the central bank reform
shock, it takes five quarters for the exchange rate to begin to appreciate,
with the highest effect around 0.5% in period 20. We confirm the evidence
of the delayed overshooting puzzle of the effect of monetary policy on the
exchange rate, in which shock monetary policy requires time to appreciate
the exchange rate, as prior studies such as Eichenbaum and Evans (1995)
and Scholl and Uhlig (2008) demonstrated.
The response of consumption to one standard deviation shock of the ex-
change rate is positive and reaches the peak of about 0.77% at quarter 6 prior
to falling in the following quarter. This implies that the depreciation ex-
change rate increases consumption which is consistent with the international
real business cycle model. Our finding is in line with Kandil (2015) who found
a positive relationship between exchange rate and consumption in emerging
economies. The positive response to exchange rate shock is also showed in
investment, where shock one standard deviation exchange rate increases 0.1%
consumption at quarter three. This result denotes that depreciation of the
exchange rate increases investment because it reduces investment cost and
hence, attracts new investment. Blonigen (1997) concludes that depreciation
improves foreign direct investment. A one percentage point to the degree of
CBI shock leads to a fall in consumption by 0.27% in the first quarter and
reaches the minimum around 0.4% at quarter 2. From period 3 onwards,
the impulse response turns back and reaches the initial value at quarter 17.
11This finding could possibly initiate greater CBI followed by the tightening
of monetary policy, for instance, reducing the supply of money to decrease
inflation, although it also reduces private consumption. Shock one standard
deviation in CBI increases investment and reaches the peak of around 0.8%
in quarter 8. Subsequently, the impact falls to the initial value at period
18. This result is in line with our expectation that a higher degree of CBI
attracts investment.
5.2. Sub-sample analysis
We then divide the sample of countries into three groups to make our sub-
samples poolable. The first group include Guatemala, Morocco, Paraguay,
Thailand, Malaysia and Trinidad and Tobago. The second group covers 5
countries: Tunisia, Uruguay, Pakistan, the Philippines and Egypt. The third
group consists of 15 countries: Mexico, Honduras, Mauritania, Suriname,
Argentina, Ghana, Kenya, South Africa, Turkey, Zambia, Bolivia, Ethiopia,
Nicaragua, Venezuela and Indonesia.
First, we perform a Panel VAR regression for subsample group 1. To
check the presence of heterogeneity across countries, we apply the Chow and
Roy-Zellner tests for pooling assumption. The results of the poolability test
show that we do not reject the null hypothesis. This signifies that the panel
is poolable, and there is no heterogeneity among the countries sample.
We can now present the impulse response function for group 1. Initially,
we can see that the effect of central bank reform shock is negative in relation
to the exchange rate. This result is in line with our expectation that higher
CBI affects the appreciation of the exchange rate. In response to one positive
standard deviation shock to the degree of CBI, the exchange rate level appre-
ciates by 0.2% in the third quarter. From period four onwards, the response
increases to reach the initial value. Shock one standard deviation exchange
rate produces lower consumption and reaches the minimum at period four
at 0.2%, while the direction changes the following quarter. The negative re-
sponse is also shown by investment to the exchange rate shock. This result
signifies that the depreciation exchange rate leads to lower investment. Fi-
nally, we can perceive that the effect of CBI shock increases economic activity,
both consumption and investment. Consumption responds positively to CBI
shock after quarter two and increases over time. The investment response to
CBI shock is positive and the coefficient is roughly 0.03% in the first period.
The response is increasing gradually and reaches the peak, roughly 0.34% at
period 20.
12Figure 2: Impulse Responses Function Model 1 Group 1
Next, we move to the second group. After apply the panel VAR regres-
sion, we check the presence of heterogeneity across countries by applying
the Chow and Roy-Zellner tests for pooling assumption. The results of the
poolability tests demonstrate that we do not reject the null hypothesis. This
implies that the panel is poolable and there is no heterogeneity among the
countries sample.
Figure 3 describes that shock one standard deviation of CBI causes an
increase in the exchange rate and reaches the peak, approximately 0.4% at
period 8, after that the response fall to the original level. Our result reveals
that greater CBI produces depreciation in the exchange rate. This confirms
the existence of the exchange rate puzzle for this group. The negative re-
sponse of consumption to one standard deviation shock of the exchange rate
begins in period 2 and reaches the minimum at period 6, about 0.55%. From
quarter 7 onwards, the response climbs to reach the initial value. In con-
trast, the investment response to one standard deviation shock of CBI is
negative, from the first period and reaches the trough by 0.46% at period
135. Subsequently, the response rises and reaches the initial value. This ev-
idence implies that the depreciation (appreciation) exchange rate leads to
lower (higher) consumption and investment. Finally, we analyse the effect of
CBI on economic activity via the exchange rate. In response to one-standard
deviation shock of CBI, the consumption response is negative for all periods.
This may well be caused by the strong effect of the depreciation exchange
rate due to CBI and therefore, generates higher imported goods. Therefore,
private consumption drops. Conversely, shock positive CBI leads to increased
investment until period 6. This implies that higher CBI together with the
depreciation exchange rate attracts investors in the short-run.
Figure 3: Impulse Responses Function Model 1 Group 2
Next, we move to the last group. We apply the Panel VAR regression on
the model, then verify the presence of heterogeneity across countries using
the Chow and Roy-Zellner tests for pooling assumption. The results of the
poolability tests show that we reject the null hypothesis. This result denotes
that the panel is not poolable and there is heterogeneity among the countries
sample. In this part, we apply the mean-group estimation by averaging the
14coefficient for the 15 countries sample.
Figure 4 presents the result of IRF’s using the mean-group estimation
for 15 countries. Shock a one-unit innovation to the degree of CBI on the
exchange rate is positive at the first 3-period and reaches the peak at 0.15%
in period 3. Subsequently, the response falls to the initial value at period
6 and remains negative until the end of period 20. This finding signifies
that CBI creates depreciation in the short-run but generates appreciation in
the long-run. In this group, we show the delayed overshooting puzzle of the
exchange rate due to changes in monetary policy. The shock a one-unit in-
novation exchange rate increases consumption and reaches the peak of about
1.8% at period 6. This group demonstrates that the depreciation exchange
rate affects higher real economic activity. These findings confirm the inter-
national business cycle model prediction of the positive relationship between
the exchange rate and consumption. Shock a one-unit innovation exchange
rate affects higher investment, roughly 0.9% at period 6. This indicates that
the depreciation exchange rate generates lower cost for investment; thus, it
increases the investment. The response of consumption to CBI shock is neg-
ative, reaching the minimum at period 4. After that, the response remains
steady. This negative effect might be caused by the strong influence of the
depreciation exchange rate. The negative investment response to CBI shock
is shown in the first 4 periods. After periods 4 to 8, the investment response
to CBI shock is positive. From period 9 onwards, the investment response is
negative. We find investment response fluctuation due to higher CBI.
5.3. Comparison sub-sample group
In this section, we analyse the interaction between CBI, exchange rate
and economic activity in various sub-samples: full sample, group 1, group 2
and group 3. The results show how these relationships are influenced by a
diversity of factors such as heterogeneity of economic sectors, degree of CBI,
inflation, exchange rate regime and capital mobility (Lane, 2001).
Figure 5 reveals the impulse response of the exchange rate to the shock
of CBI in three different groups samples and the average for the full sam-
ple. Interestingly, the response to the shock varies in each group. From
the graph, it is apparent that a one-unit shock in CBI negatively affects the
exchange rate in the first group. However, the positive effect is shown for
groups two, three and also for the full sample countries, even though after a
certain period the effect is negative. The magnitude of the maximum impact
of one percentage point change in the degree of CBI on the exchange rate
15Figure 4: Impulse Responses Function Model 1 Group 3
varies between -0.2% and 0.4%, averaging around 0.1% across the country
examined. The maximum impact is achieved at period 2 for the full sample,
group 1 and group 3; however, for group 2 is in period 8.
We discuss various economic features of developing countries to deter-
mine the main factors that can describe the different results among groups.
First, we apply the exchange rate classification measured by Reinhart and
Rogoff (2004) which is presented in ?? row 5 in the Appendix. A higher
number means a more flexible exchange rate arrangement. The three groups
have different degrees of exchange rate flexibility. Group 1 has the lowest
exchange rate flexibility; the average group members in group 1 are applying
de facto crawling peg. Group 2 has moderate exchange rate flexibility; a
number of countries are using the crawling band exchange rate arrangement.
Group 3 is the highest degree; certain countries adopt the manage floating
exchange rate. Theoretically, the peg exchange rate arrangement may result
in exchange rate puzzling (Kim and Lim, 2016). In the 1990s, most devel-
oping countries applied the peg exchange rate system to limit speculation,
provide a stable system for traders and investors, and prevent market adjust-
ments when a currency undervalued. However, this system creates an unreal
exchange rate or exchange rate manipulation. Therefore, there is a gap be-
tween the official and unofficial exchange rate. Independent central banks
commonly followed by changing the exchange rate system from the peg to
16the floating exchange rate system. In the short-run, the exchange rate will
adjust to the market value. Therefore, the exchange rate will be depreci-
ated as it is undervalued. However, in the long-run, central bank reforms
could lower the exchange rate depreciation. This implies that under CBI,
monetary policy is predictable, generating lower exchange rate uncertainty.
This finding is relevant with Cermeño et al. (2010), who concluded that cen-
tral bank reforms are associated with reduced exchange rate uncertainty and
lower average depreciation in the long-run.
According to Mundell (1963), the link between monetary policy and the
exchange rate based on perfect capital mobility. When capital mobility is
highly controlled, the increase in CBI degree may not effect capital flows
and the exchange rate. ?? row 6 in the Appendix describes the capital con-
trol restriction developed by Fernández et al. (2016). We establish that the
degree of capital control restrictions are different for the three subgroups, re-
sulting in different monetary transmission effects on the exchange rate. The
results for group 1, which is the lowest capital control restriction, are com-
patible with theoretical expectations that higher CBI leads to lower inflation
expectation and hence, will appreciate the exchange rate. The ”overshoot-
ing” theory developed by Dornbusch (1976) predicts that in the long-run,
the exchange rate should initially overshoot to respond to monetary policy
shock. Thus, an overshooting exchange rate occurred in this group. Previous
empirical results support this finding, for instance, Kim and Roubini (2000),
Faust and Rogers (2003) and Bjørnland (2009), who ascertained an appre-
ciation exchange rate to monetary policy shocks. In contrast, the exchange
rate puzzle is demonstrated by group 2, where monetary policy shock affects
exchange rate depreciation. In this group, capital mobility restriction is high.
The previous result presented by Grilli and Roubini (1995) strengthen this
finding. According to Kim and Lim (2016), exchange rate puzzle occurs in
countries with strong restricted capital mobility. Hence, the change in mon-
etary policy may not influence the exchange rate. The result for group 3 and
the full sample group add the existence of the ”delayed overshooting” puzzle
previously documented by Eichenbaum and Evans (1995) and Scholl and Uh-
lig (2008). Group 3 has moderate capital mobility restriction. Eichenbaum
and Evans (1995) found that the exchange rate overshoots is the long-run
effect of monetary policy shock, though it occurs one to three years after
the shocks. These delays can be caused by ”forward discount bias” puzzle,
conditional on monetary policy shocks.
The left side of Figure 6 shows the impulse responses of consumption to
17Figure 5: Impulse Response Function of Exchange Rate to CBI Shock
a one positive unit shock in the exchange rate. The effects of the exchange
rate on consumption are positive in the full sample and group 3. This sug-
gests that currency depreciation generates higher consumption which sup-
ports the international real business cycle model. This finding is consistent
with the international real business cycle model which predicts the positive
relationship between domestic consumption and the depreciation exchange
rate. The depreciation exchange rate generates less expensive exports, while
imports are more expensive. Based on competitiveness, foreign demand for
exports goods and services will increase and domestic demand for imports
will decrease; thus, shifting consumption to domestic goods and services.
As the net external demand increases, the increases in aggregate demand
will have a positive effect on output and create jobs. Increasing income
and employment generates increased consumption. Past empirical studies
such as Kandil (2015), found that the depreciation exchange rate initiates
increased consumption growth in developing countries, due to increases in
the domestic prices of import goods and services. Conversely, for groups 1
and 2, consumption increases due to the appreciation exchange rate. Ap-
preciation of the exchange rate decreases the cost of tradables and nontrad-
ables goods and, therefore, increases consumption growth. The combined
effect will depend on the elasticity of consumers’ substitution between trad-
ables and non-tradables (Kandil and Mirzaie, 2006). This result confirms the
18Figure 6: Impulse Response Function of Economic Activity to Exchange Rate Shock
Backus-Smith puzzle of a negative relationship between exchange rate and
consumption (Backus and Smith, 1993). According to Kollmann (2012), the
consumption-real exchange rate anomaly might be due not to the underde-
velopment of international financial markets. Additional empirical evidence
such as Devereux et al. (2012), determined a negative relationship between
exchange rate and consumption.
The right side of Figure 6 illustrates the impulse responses of the invest-
ment to a one-unit shock in the exchange rate depreciation can create an
increase in investment as the marginal profit from an additional unit of capi-
tal is likely to increase as future foreign sales rise for group 3 and the average
all countries. Conversely, for groups 1 and 2, the appreciation exchange rate
leads to higher investment. Depreciation produces higher price for imported
capital, whilst inputs can reduce profits and, in turn, lower investment. The
overall impact of the exchange rate changes on investment hinges on which
of these forces dominates (Landon and Smith, 2009). Depreciation of the ex-
change rate boosts investment, reducing the cost of investment in domestic
countries. Increases in investment are also consistent with higher demand for
goods and services because competition increases. This indicates the more
dominant supply-side effect via the increase of competitiveness channel, par-
ticularly in the tradable sector.
Currency depreciation also has a positive implication for investment by
means of relative wage channels. Depreciation reduces domestic’s wages and
production costs compared to those of its foreign counterparts. Therefore,
it increases the overall rate of return on foreign investment in the domes-
tic country. Previous studies, such as Blonigen (1997), argue that foreign
19exchange rate depreciation will lead to enhanced foreign direct investment
(FDI) into the foreign economy. This result is also in lines with the Mundell-
Fleming model for open economies which describes that the depreciation
of domestic currency generates greater investment, seeing that it produces
cheaper domestic goods and thus, more competitive in international markets.
The principal factor related to a company’s investment decisions are price
competitiveness (Brito et al., 2018). In contrast, appreciation of domestic
currency should stimulate investment for countries that depend on imported
capital goods, as foreign capital goods are cheaper (Alejandro, 1963). Specific
studies show that currency depreciations (appreciations) are associated with
a contraction (expansion) in investment (Landon and Smith, 2009; Goldberg,
1993; Campa and Goldberg, 2005). Related to these inconclusive findings
are the financial effects of exchange rate changes, including those operating
through adjustments in balance sheets, which are reviewed below. Further-
more, there is also extensive theoretical literature looking at how linkages
between exchange rate and output can depend on exchange rate regimes
(Uribe, 1997) and (Mendoza and Uribe, 1997).
The different effect of the exchange rate on economic activity might be
caused by average level of inflation. Regarding ?? row 3 in Appendix; Group
1 comprises a low average inflation rate, group 2 has a moderate inflation rate
and group 3 consists of a high inflation rate, while the average for all coun-
tries is close to the average of group 3. For low and moderate inflation rate,
the appreciation (depreciation) exchange rate leads to higher (lower) con-
sumption and investment. The positive effect is higher for the low inflation
group. Conversely, in countries with high inflation, depreciation (apprecia-
tion) causes higher (lower) consumption and investment.
Finally, we determine that the direct link from CBI to consumption and
investment in Figure 7. The reaction of consumption due to changes in CBI
is negative in the first quarter for all groups. Only after a lag of approxi-
mately 3 quarters for group 1 does the impact become positive, though for
the other groups the response is negative until period 20. CBI has a posi-
tive effect on consumption only in countries with low inflation, given that a
greater degree of CBI gives central banks freedom to set monetary policy, for
instance lowering interest rates and increasing the money supply. Thus, these
expansionary monetary policies create higher consumption. In contrast, in
group 2 and group 3, which have high inflation rates, the central bank fo-
cuses more on lowering price by setting contractionary monetary policy such
as, tight money policy. Tightening the supply of money lowers private con-
20sumption. Generally, the average for all countries sample is that higher CBI
creates lower consumption, seeing as regularly, they have high inflation and
the target of the central bank is to reduce prices.
Figure 7: Impulse Response Function of Economic Activity to CBI Shock
The reason for different consumption responses to CBI is the sensitivity
of imported goods (Carriere-Swallow et al., 2017). One major problem in
developing countries is high inflation owing to lack of supply. One possi-
ble solution is importing goods to lower the price because price in foreign
countries is lower than domestically. For group 1, greater CBI causes an
appreciation of the exchange rate, making imported goods cheaper. Hence
households can consume more. In contrast, CBI affects the depreciation ex-
change rate in group 2, group 3 and in the full sample. This depreciation
generates more imported goods, therefore, reducing consumption.
Figure 7 also presents the investment responses to CBI shock in various
groups. We can see the positive effect of CBI on investment, seeing that
increasing the degree of CBI as a signal to combat inflation, display good
governance institution and consequently, will attract investment. For group
2 and group 3, the responses become negative after six and 8 quarters. Here,
the central bank could focus on reducing inflation by raising the interest rate
and thus, discouraging investment. In group one, increasing the degree of
CBI is also a signal for the implementation of structural economic reforms
(Lavezzolo and de Estudios Avanzados en Ciencias Sociales, 2006). This will
create good opportunities for investors and moreover, promote investment.
Generally, the inverse relationship of CBI to real economic activity can be
explained by there the trade-off between inflation and output. Because the
monetary authorities prefer price stability; thus, they less concern regarding
21the real activity. These findings are related to Barro and Gordon (1983) who
stated that there is always a trade-off between credibility and flexibility, given
that the difference between economic activity (growth and unemployment)
and inflation can be viewed as the main difference between discretion and
rules. Our findings are supported by previous studies, for example Alesina
and Summers (1993), Cukierman et al. (1992), Froyen and Waud (1995) and
Parkin (2013), who found no association between CBI and real economic
activity.
6. Conclusion
This chapter provides empirical analysis of CBI and economic activity
via three different asset prices: exchange rate, stock price and bond yield. It
begins with the methodological design of our empirical analysis. Initially, the
panel VAR is applied for four different models. Panel VAR is selected for the
reason that it is the most appropriate method and it has the ability to treat
all variables as endogenous. We suspect the appearance of heterogeneity
among cross-sections; hence, we verify the pooling assumption of the panel.
We find the heterogeneity of the sample; therefore, we apply mean-group
estimation developed by Pesaran and Smith (1995) for panel VAR. We also
split our sample into two and three groups so our subsamples are poolable.
First, we examine the interaction between CBI, exchange rate and eco-
nomic activity. Our results using panel VAR indicate the model contains
heterogeneity across countries. We then apply mean-group panel VAR by
averaging the individual VAR for all samples. We establish that the nega-
tive response of the exchange rate to CBI shock is delayed; it takes around
5 quarters to appreciate exchange rate. The effect of the exchange rate on
both consumption and investment is positive, which implies that deprecia-
tion in the exchange rate creates higher consumption and investment. We
determine that CBI has a contradictory effect on real activity, a negative
effect on consumption but a positive effect on investment. We then split our
sample into three groups to create homogeneous subsample. Results show
that CBI produces appreciation of the exchange rate for group 1 but depre-
ciation of the exchange rate for group 2. For group 3, CBI strengthens the
exchange rate after 6 quarters of the shock. Meanwhile, the response of eco-
nomic activity to the exchange rate also vary for all three groups. We find
a negative response for both consumption and investment to the exchange
rate for groups 1 and 2. This means that appreciation of the exchange rate
22generates higher consumption and investment. However, for group 3 we find
the opposite result, which is depreciation will increase both consumption and
investment. Finally, the CBI link to real activity via the exchange rate chan-
nel can be explained as follows. Higher CBI degree increases consumption
and investment only occurs in countries with low inflation countries (group
1). In contrast, CBI causes lower consumption and investment for countries
with moderate and high inflation (groups 2 and 3).
Acknowledgement
The authors would like to acknowledge financial support from the Indone-
sian Endowment Fund for Education (LPDP), Ministry of Finance Republic
of Indonesia; and the Directorate General of Higher Education (DIKTI),
Ministry of Research, Technology and Higher Education Republic of Indone-
sia.
References
Acemoglu, D., Johnson, S., Querubin, P., Robinson, J.A., 2008. When Does Policy
Reform Work? The Case of Central Bank Independence. Working Paper 14033.
National Bureau of Economic Research. URL: http://www.nber.org/papers/
w14033, doi:10.3386/w14033.
Agoba, A.M., Abor, J., Osei, K.A., Sa-Aadu, J., 2017. Central bank inde-
pendence and inflation in africa: The role of financial systems and institu-
tional quality. Central Bank Review 17, 131 – 146. URL: http://www.
sciencedirect.com/science/article/pii/S1303070117300550, doi:https:
//doi.org/10.1016/j.cbrev.2017.11.001.
Alejandro, C.F.D., 1963. A note on the impact of devaluation and the redistributive
effect. Journal of Political Economy 71, 577–580.
Alesina, A., Summers, L.H., 1993. Central bank independence and macroeco-
nomic performance: Some comparative evidence. Journal of Money, Credit and
Banking 25, 151–162. URL: http://www.jstor.org/stable/2077833.
Backus, D.K., Smith, G.W., 1993. Consumption and real exchange rates in dy-
namic economies with non-traded goods. Journal of International Economics
35, 297 – 316. URL: http://www.sciencedirect.com/science/article/pii/
002219969390021O, doi:https://doi.org/10.1016/0022-1996(93)90021-O.
23Baltagi, B., 2008. Econometric analysis of panel data. John Wiley & Sons.
Barro, R., Gordon, D., 1983. Rules, discretion and reputation in a model of
monetary policy. Journal of Monetary Economics 12, 101–121. URL: https:
//EconPapers.repec.org/RePEc:eee:moneco:v:12:y:1983:i:1:p:101-121.
Bearce, D.H., 2008. Not complements, but substitutes: fixed exchange rate com-
mitments, central bank independence, and external currency stability. Interna-
tional Studies Quarterly 52, 807–824.
Bjørnland, H.C., 2009. Monetary policy and exchange rate overshoot-
ing: Dornbusch was right after all. Journal of International Economics
79, 64 – 77. URL: http://www.sciencedirect.com/science/article/
pii/S0022199609000841, doi:https://doi.org/10.1016/j.jinteco.2009.
06.003.
Blonigen, B.A., 1997. Firm-specific assets and the link between exchange rates
and foreign direct investment. The American Economic Review 87, 447–465.
URL: http://www.jstor.org/stable/2951354.
Brito, S., Magud, M.N.E., Sosa, M.S., 2018. Real Exchange Rates, Economic
Complexity, and Investment. International Monetary Fund.
Campa, J.M., Goldberg, L.S., 2005. Exchange Rate Pass-Through into Import
Prices. The Review of Economics and Statistics 87, 679–690. URL: https:
//ideas.repec.org/a/tpr/restat/v87y2005i4p679-690.html.
Canova, F., Ciccarelli, M., 2013. Panel vector autoregressive models: A survey.
Advances in Econometrics 31. doi:10.1108/S0731-9053(2013)0000031006.
Carriere-Swallow, M.Y., Gruss, B., Magud, M.N.E., Valencia, M.F., 2017. Mone-
tary policy credibility and exchange rate pass-through. International Monetary
Fund.
Cermeño, R., Grier, R., Grier, K., 2010. Elections, exchange rates and reform in
latin america. Journal of Development Economics 92, 166–174.
Claessens, S., Kose, M.A., 2017. Asset prices and macroeconomic outcomes: a
survey. The World Bank.
Cukierman, A., Webb, S.B., Neyapti, B., 1992. Measuring the independence of
central banks and its effect on policy outcomes. The World Bank Economic
Review 6, 353–398. URL: http://www.jstor.org/stable/3989977.
24Devereux, M.B., Smith, G.W., Yetman, J., 2012. Consumption and real ex-
change rates in professional forecasts. Journal of International Economics
86, 33 – 42. URL: http://www.sciencedirect.com/science/article/
pii/S0022199611001127, doi:https://doi.org/10.1016/j.jinteco.2011.
08.014.
Dornbusch, R., 1976. Expectations and exchange rate dynamics. Journal of Polit-
ical Economy 84, 1161–1176. URL: http://www.jstor.org/stable/1831272.
Eichenbaum, M., Evans, C.L., 1995. Some empirical evidence on the effects of
shocks to monetary policy on exchange rates. The Quarterly Journal of Eco-
nomics 110, 975–1009.
Eichler, S., Littke, H.C., 2018. Central bank transparency and the volatil-
ity of exchange rates. Journal of International Money and Finance
89, 23 – 49. URL: http://www.sciencedirect.com/science/article/
pii/S0261560618300548, doi:https://doi.org/10.1016/j.jimonfin.2018.
07.006.
Faust, J., Rogers, J.H., 2003. Monetary policy’s role in exchange rate behav-
ior. Journal of Monetary Economics 50, 1403 – 1424. URL: http://www.
sciencedirect.com/science/article/pii/S0304393203000904, doi:https:
//doi.org/10.1016/j.jmoneco.2003.08.003.
Fernández, A., Klein, M.W., Rebucci, A., Schindler, M., Uribe, M., 2016. Capital
control measures: A new dataset. IMF Economic Review 64, 548–574. URL:
https://doi.org/10.1057/imfer.2016.11, doi:10.1057/imfer.2016.11.
Froyen, R.T., Waud, R.N., 1995. Central bank independence and the
output-inflation tradeoff. Journal of Economics and Business 47, 137
– 149. URL: http://www.sciencedirect.com/science/article/pii/
014861959400042C, doi:https://doi.org/10.1016/0148-6195(94)00042-C.
central Banking at a Crossroads.
Garriga, A.C., 2016. Central bank independence in the world: A
new data set. International Interactions 42, 849–868. URL: https:
//doi.org/10.1080/03050629.2016.1188813, doi:10.1080/03050629.2016.
1188813, arXiv:https://doi.org/10.1080/03050629.2016.1188813.
Gerlach, S., Assenmacher-Wesche, K., 2008. Financial structure and the impact
of monetary policy on asset prices. CFS Working Paper Series 2008/30. Center
for Financial Studies (CFS). URL: https://ideas.repec.org/p/zbw/cfswop/
200830.html.
25Goldberg, L.S., 1993. Exchange Rates and Investment in United States Industry.
The Review of Economics and Statistics 75, 575–588. URL: https://ideas.
repec.org/a/tpr/restat/v75y1993i4p575-88.html.
Grilli, V., Roubini, N., 1995. Liquidity and Exchange Rates: Puzzling Evidence
from the G-7 Countries. Technical Report. New York University, Leonard N.
Stern School of Business, Department of Economics.
Kandil, M., 2015. On the benefits of nominal appreciations: Contrasting evi-
dence across developed and developing countries. Borsa Istanbul Review 15,
223 – 236. URL: http://www.sciencedirect.com/science/article/pii/
S2214845015000289, doi:https://doi.org/10.1016/j.bir.2015.06.003.
Kandil, M., Mirzaie, I.A., 2006. Consumption and macroeconomic policies: Theory
and evidence from developing countries. J. Int. Trade & Economic Development
15, 469–491.
Kim, S., Lim, K., 2016. Effects of Monetary Policy Shocks on Exchange Rate in
Emerging Countries. Working Papers 192016. Hong Kong Institute for Monetary
Research. URL: https://ideas.repec.org/p/hkm/wpaper/192016.html.
Kim, S., Roubini, N., 2000. Exchange rate anomalies in the industrial
countries: A solution with a structural var approach. Journal of Mon-
etary Economics 45, 561 – 586. URL: http://www.sciencedirect.com/
science/article/pii/S0304393200000106, doi:https://doi.org/10.1016/
S0304-3932(00)00010-6.
Kollmann, R., 2012. Limited asset market participation and the consumption-
real exchange rate anomaly. Canadian Journal of Economics/Revue canadienne
d’économique 45, 566–584.
Landon, S., Smith, C., 2009. Investment and the exchange rate: Short run and
long run aggregate and sector-level estimates. Journal of International Money
and Finance 28, 813–835. URL: https://EconPapers.repec.org/RePEc:eee:
jimfin:v:28:y:2009:i:5:p:813-835.
Lane, P.R., 2001. The new open economy macroeconomics: a survey. Journal of In-
ternational Economics 54, 235 – 266. URL: http://www.sciencedirect.com/
science/article/pii/S0022199600000738, doi:https://doi.org/10.1016/
S0022-1996(00)00073-8.
26Lavezzolo, S., de Estudios Avanzados en Ciencias Sociales, C., 2006. Central
Bank Independence in Developing Countries: A Signaling Mechanism? Estu-
dios (Centro de Estudios Avanzados en Ciencias Sociales (Madrid))), Centro de
Estudios Avanzados en Ciencias Sociales. Instituto Juan March de Estudios e In-
vestigaciones. URL: https://books.google.co.uk/books?id=I58UAQAAIAAJ.
Mendoza, E., Uribe, M., 1997. The Syndrome of Exchange-Rate-Based Stabi-
lizations and the Uncertain Duration of Currency Pegs. Working Papers 97-30.
Duke University, Department of Economics. URL: https://ideas.repec.org/
p/duk/dukeec/97-30.html.
Moser, C., Dreher, A., 2010. Do markets care about central bank governor changes?
evidence from emerging markets. Journal of Money, Credit and Banking 42,
1589–1612.
Mundell, R.A., 1963. Capital mobility and stabilization policy under fixed and flex-
ible exchange rates. Canadian Journal of Economics and Political Science/Revue
canadienne de economiques et science politique 29, 475–485.
Parkin, M., 2013. The Effects of Central Bank Independence and Inflation Tar-
geting on Macroeconomic Performance: Evidence from Natural Experiments.
University of Western Ontario, Economic Policy Research Institute Working
Papers 20133. University of Western Ontario, Economic Policy Research Insti-
tute. URL: https://EconPapers.repec.org/RePEc:uwo:epuwoc:20133.
Pesaran, M., Smith, R., 1995. Estimating long-run relationships from dynamic het-
erogeneous panels. Journal of Econometrics 68, 79 – 113. URL: http://www.
sciencedirect.com/science/article/pii/030440769401644F, doi:https:
//doi.org/10.1016/0304-4076(94)01644-F.
Reinhart, C.M., Rogoff, K.S., 2004. The Modern History of Exchange Rate Ar-
rangements: A Reinterpretation. The Quarterly Journal of Economics 119, 1–48.
URL: https://ideas.repec.org/a/oup/qjecon/v119y2004i1p1-48..html.
Rogoff, K., 1985. The optimal degree of commitment to an intermediate monetary
target. The Quarterly Journal of Economics 100, 1169–1189. URL: http:
//www.jstor.org/stable/1885679.
Sa, F., Towbin, P., Wieladek, T., 2011. Low interest rates and housing booms:
The role of capital inflows, monetary policy and financial innovation. SSRN
Electronic Journal doi:10.2139/ssrn.1765853.
27Sanchez, M., 2008. The link between interest rates and exchange rates: do con-
tractionary depreciations make a difference? International Economic Journal
22, 43–61.
Scholl, A., Uhlig, H., 2008. New evidence on the puzzles: Results from agnostic
identification on monetary policy and exchange rates. Journal of International
Economics 76, 1–13.
Uribe, M., 1997. Exchange-rate-based inflation stabilization: The initial real effects
of credible plans. Journal of Monetary Economics 39, 197 – 221. URL: http:
//www.sciencedirect.com/science/article/pii/S0304393297000184,
doi:https://doi.org/10.1016/S0304-3932(97)00018-4.
28You can also read
